U.S. labor market concentration, competition, and worker bargaining power as employment trends shift from manufacturing to services
Labor market concentration and competition in the United States has received renewed and well-deserved attention over the past several years. Competitive labor markets are critical engines of economic growth, worker well-being, and equality. Like in the concentration of sales data, however, it is well-documented that concentration levels in the U.S. labor sector have been rising at the national level.
A new working paper, titled “Local and National Concentration Trends in Jobs and Sales: The Role of Structural Transformation,” confirms these national trends. Its co-authors—David Autor at the Massachusetts Institute of Technology, Christina Patterson at the University of Chicago, and John Van Reenan at the London School of Economics—find that concentration levels across all sectors—manufacturing, retail trade, wholesale trade, services, utilities and transportation, and finance—individually and cumulatively have risen sharply.
They calculate that concentration in sales, as measured by a weighted average of the Herfindahl-Hirschman Index, or HHI—a widely used market-concentration measure that is calculated by summing the squares of each firm’s share in a given market—is up 53 percent and concentration in employment is up by 68 percent in 2017, compared to 1992. To put this in context, a market that goes from five equal-size companies to only three equal-size companies will experience a 63 percent increase in its HHI.
More significantly, the paper also looks behind these national concentration figures to examine whether local concentration trends in job markets and sales track these national trends. These local trends are arguably more important from a competition perspective because people prefer to shop and work near where they live, making competition in local markets most directly relevant to people’s day-to-day lives and livelihoods.
The three co-authors find that while local sales concentration trends track national trends, local job markets have become less concentrated even as national labor concentration has trended upward. Interestingly, though, the downward trend in local job market concentration is being driven by sectoral shifts, not by de-concentration of job markets in any sector.
The authors identify the reallocation of economic activity away from the manufacturing sector and toward the service sector as key to understanding the net decline in concentration in local labor markets. They find that manufacturing is much more highly concentrated than services, even though concentration in both industries has risen over their period of analysis.
For sales, the sectoral shift is not enough to offset the rising concentration trend, and local sales concentration still rises. In employment, however, they conclude that “the reallocation of employment toward less concentrated industry-county cells more than fully offsetsthe rise in within industry-county employment concentration, leading to a net fall in local employment concentration.” Without the reallocation, they estimate that employment concentration would have risen by 9 percent instead of falling by 6.9 percent.
Autor, Patterson, and Van Reenan explain the disparate results between employment and sales concentrations by observing that productivity—output per worker—generally rises faster in manufacturing that in other sectors, meaning that a reallocation from manufacturing to other sectors tends to be larger than for sales. “All else equal, sectoral reallocation from manufacturing to services reduces both local sales and employment concentration,” they write. “But the de-concentrating effect is larger for employment than sales because the extent of reallocation is greater.”
The co-authors use a series of counterfactuals to illustrate the fact and the magnitude of this effect. They also look at whether geographic shifts can explain the observed differences, finding that controlling for geographic shifts only modestly impacts trends in employment concentration and has no effect on sales concentration.
The implications of these findings are nuanced but significant. Regarding sales, the co-authors note that increased concentration is likely driven by the rise of chain stores, which may be more efficient and offer better variety. Regarding employment, workers with good sectoral mobility probably enjoy greater competition, and those with limited sectoral mobility have more limited options. But, importantly, that increased overall competitiveness comes not from employment markets becoming more competitive—they are not—but rather from workers shifting from less competitive sectors (manufacturing) into relatively more competitive sectors (services).
It remains to be seen whether trends in increasing local employment competitiveness will hold up over time, but there are good reasons to suspect they may not. First, at some point, the shift of workers from manufacturing to services will slow down or even stop, as the pool of manufacturing workers shrinks. Indeed, manufacturing’s share of employment has already fallen by almost 10 percent from 1992 to 2017. As noted earlier, without the effect of this shift the authors estimate employment concentration would have risen by 9 percent instead of falling by 6.9 percent over the studied period.
Second, employment in the service sector is more competitive than in manufacturing, yet both sectors are trending toward increased concentration. Combining these two effects over time may see the impact of the shift shrink while an increasingly large pool of service workers see concentration in that sector increase.
The reallocation of workers from manufacturing to services also has significant implications for the balance of bargaining power between workers and employers, highlighting the close-but-complex relationship between bargaining power and competition. While employment markets in manufacturing are more concentrated than in services, manufacturing workers have much higher rates of union membership and a more organized labor force.
As a result, employees in manufacturing markets, even though concentrated, have historically enjoyed relatively high wages and a greater share of the economic surplus generated in these markets than their counterparts in the service sector. Accordingly, the reallocation of workers from manufacturing to services might simultaneously lead to decreased concentration in local labor markets and decreased worker bargaining power—and attendant lower wages and poorer working conditions.
At the same time, the reallocation of workers to the service sector also may be why there is new interest in unionization within that sector. A resurgent labor movement has met opposition from service-sector employers, such as Starbucks Corp., and led to calls for increased government support for unionization efforts, particularly in the service sector.
As U.S. antitrust agencies and competition policymakers begin to take competition in labor markets more seriously, they will also have to grapple with the important role that bargaining power plays in labor markets. While the recently proposed draft Federal Trade Commission-U.S. Department of Justice Merger Guidelines for Public Comment’s direct address of labor markets and labor market harms is an encouraging indication that the agencies expect sustained focus on labor issues, they give little indication of how the role of bargaining power in labor markets will be addressed.
Equitable growth in competitive labor markets will require more than just antitrust enforcement, as worker organization also plays a critical role. This study by Autor, Patterson, and Van Reenan drives home the complexity of understanding U.S. labor market dynamics at both the national and the local level, and the need to pursue multifaceted policy solutions to promote efficient and equitable markets for workers.